Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Cholera, caused by a Gram-negative bacterium, Vibrio cholerae, is commonly considered to be dependent on a cAMP-mediated active secretory mechanism. Several other mediators have, however, been implicated in the mediation of CT-induced intestinal fluid secretion. In 1970, an increase in 5-HT levels in the blood of choleraic rabbits was demonstrated (Bhide et al., 1970). CT administered into the duodenums of rabbits caused severe degranulation of enterochromaffin cells as revealed by electron microscopy (Osaka et al., 1975). Since these observations, substantial evidence has accumulated to prove the involvement of 5-HT in the genesis of CT-induced fluid and electrolyte secretion (Cassuto et al., 1982; Nilsson et al., 1983; Beubler et al., 1989, Turvill et al., 1998). It has been furthermore suggested that CT may cause diarrhea by stimulating prostaglandin (PG) synthesis (Beubler et al., 1989). This concept is supported by the observation that CT is apparently associated with increased local PG synthesis (Bedwani et al., 1975; Speelmann et al., 1985) and that cyclooxygenase inhibitors impair the secretory effect of CT (Wald et al., 1977). The finding that indomethacin in some studies inhibits CT-induced secretion but not mucosal cAMP accumulation (Wald et al., 1977; Beubler et al., 1986) favors the notion that PGs may play a primary role in the secretory mechanism. On the other hand, PGE₂ has been shown to be an important intermediate in the transduction mechanism, which leads to 5-HT-induced intestinal secretion (Beubler et al., 1986). From these experiments, a hypothesis was proposed that CT stimulates an apical receptor on the enterochromaffin cells and that serotonin released by the stimulus may cause PG formation, which mediates the diarrheogenic action of CT. However, the isoenzymes of cyclooxygenase, COX-1, or COX-2 involved in the relevant PG synthesis has not been determined. COX-1 is considered the constitutive form, being expressed in most tissues and cells (O'Neill and Ford-Hutchinson, 1993) and COX-2, the inducible form that is increased by inflammatory stimuli (Kujubu et al., 1991) and mitogenic stimuli (Raz et al., 1989) as well as in tumors (Eberhart et al., 1994) and ulcerated gastric tissue (Mizuno et al., 1997; Schmassmann et al., 1998). However, the constitutive expression of COX-2 was demonstrated in several tissues, including the brain, kidney, female reproductive tract, and other organs. Recently, constitutive COX-2 mRNA was identified in rat jejunal mucosa (Beubler et al., 1999) and constitutive COX-2 protein was demonstrated in mouse colon (MacNaughton et al., 2000). The aim of the present study was to elucidate the role of the different isoforms of cyclooxygenase, COX-1 and COX-2, in CT-induced fluid secretion by using selective COX-1 and COX-2 inhibitors, and by determining mRNA for COX-1 and COX-2, and COX-2 protein levels in mucosal scrapings of the rat jejunum.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Preparation of Animals. Female Sprague-Dawley rats (180 ± 20 g body weight; obtained from Himberg, Institut für Labortierkunde und -genetik, Universität Wien, Medizinische Fakultät, Austria) were used in this study. They were maintained on a standard laboratory diet and deprived of food for 18 h before the experiments but had free access to water. The rats were anesthetized with sodium pentobarbitone (60 mg/kg i.p.), the abdomen was opened via a midline incision, and a polyethylene catheter (PE60) was placed in the jejunum ~5 cm distal to the flexura duodenojejunalis and fixed by ligation. The second ligation was made ~20 cm distal to the first ligation. The loop was carefully rinsed with 20 ml of body-warm saline and the distal ligation was tied off. Intact blood flow to the jejunum is maintained by this preparation. The jejunal loop was then returned to the abdominal cavity, and the whole preparation was allowed to rest for 1 h under a heat lamp to preserve body temperature. Anesthesia was maintained by s.c. injection of sodium pentobarbitone (20 mg/kg).

Determination of Net Fluid Transport. Net fluid transfer rates were determined gravimetrically 4 h after instillation of Tyrode's solution or Tyrode's solution plus CT. The catheter was removed, the proximal ligation tied off, and the jejunal loop quickly withdrawn and weighed. Net fluid transport was expressed as milliliters per gram (ml/g) wet weight of jejunum. Net absorption was indicated by a negative value and net secretion by a positive value.

Determination of Prostaglandin E₂. Prostaglandin E₂ was determined in aliquots of jejunal fluid obtained 4 h after Tyrode or Tyrode plus CT exposure. To minimize PGE₂ release due to mechanical stimulation, the gut wall was punctured with a hypodermic needle and the fluid carefully withdrawn into a 1-ml syringe. Prostaglandin E₂ was determined by radioimmunoassay as described previously (Jobke et al. 1973, Schuligoi et al. 1998) using [5,6,8,11,12,14,15(N)-³H]-PGE₂ (Amersham, Little Chalfont, UK) as tracer and synthetic PGE₂ (Sigma) as standard.

Treatment of Rats with LPS. A separate set of experiments was performed to examine the effect of Escherichia coli LPS (Sigma) on COX mRNA expression. Rats were deprived of food for 18 h before the experiments but had free access to water. Three hours after treatment with LPS (5 mg/kg i.p.) or 0.9% NaCl, rats were sacrificed with an overdose of pentobarbital. The jejunum was taken out, and scrapings of the mucosa were used for RNA extraction.

Reverse Transcription-Polymerase Chain Reaction (RT-PCR). RT-PCR was performed as described previously (Amann et al., 1999) with total RNA from the mucosa of the jejunum (scrapings were taken immediately after the determination of net fluid transfer rates and frozen in liquid nitrogen). The RNA was extracted using Trizol (Life Technologies, Lofer, Austria) and after treatment with RNase-free DNase I (Roche, Vienna, Austria) to remove contaminating DNA, RNA was purified using a Nucleo-Spin Kit (Machery and Nagel, Düren, Germany). Reverse transcription of 0.8 µg RNA was performed with avian myoblastosis virus reverse transcriptase, and an oligo(dT)₁₅ primer (Promega, Mannheim, Germany). For COX-1, COX-2 and glyceraldehyde-3 phosphate dehydrogenase (GAPDH), specific primers were used to amplify the reverse-transcribed products. For COX 1, 1 µl, for COX-2, 3 µl and for GAPDH, 1 µl of the reverse- transcribed product were used for PCR. PCR was carried out using 2.5 mM MgCl₂, 0.2 mM deoxynucleotide triphosphates (Sigma), 20 pmol primers and 0.25 U of Taq polymerase (Promega). Samples were denatured for 2 min at 95°C and specific cDNA amplified using a Stratagene RoboCycler (Stratagene, La Jolla, CA) cycling program: 95°C for 1 min, 72°C for 2 min. Final extension time was 4 min at 72°C. A total of 33 cycles were performed for COX-1 and COX-2 and 25 for GAPDH amplification. The yield of the amplified product was tested to be in the linear range for amount of input of the reverse-transcribed product and PCR cycle number and optimal conditions were chosen. The PCR products were separated by electrophoresis in an agarose gel stained with ethidium bromide. Specific primers for COX-1 and COX-2 (Beiche et al., 1998) yielded an amplification product of 447 bp and 279 bp, respectively, and for GAPDH (obtained from CLONTECH, Heidelberg, Germany) a 450-bp fragment was obtained. GAPDH was used as an internal reference to verify similar levels of RNA used for reverse transcription.

Western Immunoblot Analysis on COX-2. Tissue samples of rat jejunal mucosa were dissolved in RIPA lysis buffer (1× phosphate-buffered saline, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS, Sigma) containing protease inhibitors [100 µg/ml phenylmethylsulfonyl fluoride (Sigma), aprotinin 30 µg/ml, and 100 mM sodium orthovanadate (10 µl/ml)]. Tissue homogenates were prepared by disrupting the tissues by repeated vortexing and boiling for 10 min. Samples were centrifuged at 15,000g for 20 min at 4°C to obtain clear lysate. The supernatant was transferred to a new microcentrifuge tube, and the remaining pellet was dissolved in 500 µl of SDS-loading buffer. Protein concentration of supernatant was determined using a BCA protein determination kit (Pierce, Rockford, IL). Likewise, we determined the protein concentration before loading the pellet dissolved in SDS by performing sample dilutions of 1:2 and 1:10, then by testing with a BCA protein assay kit. Equal amounts of protein from supernatant and pellet were loaded onto an SDS-12% polyacrylamide gel and electrophoresed according to the manufacturer's instructions (Bio-Rad). After electrophoresis, samples were transferred to nitrocellulose membranes. Once transfer was complete, membranes were briefly rinsed in 1× TBS (Tris-buffered saline) and blocked in 3% gelatin [in 1× TTBS (TBS + 0.05% Tween 20)] overnight at room temperature. After blocking, the gelatin was removed and the membrane washed twice for 10 min in 1× TTBS, followed by incubation in primary antibody (1 µl/ml) in 1% gelatin (in 1× TTBS) for 2 h. The polyclonal antibodies to COX-2 were made to a region, which was conserved between mouse, rat, and human, and were obtained from Santa Cruz Biotechnology, Santa Cruz, CA (catalog number SC1745). After incubation, the membrane was rinsed twice with 1× TTBS for 10 min each followed by incubation with appropriate alkaline phosphatase (AP)-conjugated secondary antibody (diluted 1:5000) (Santa Cruz Biotechnology) in 1% gelatin for 1 h. After incubation, the membrane was rinsed three times with 1× TTBS for 10 min each followed by a final wash in 1× TBS for 10 min. After washes, AP substrate (Bio-Rad) was applied to the membrane according to the manufacturer's instructions.

Drugs. NS-398 ([N-(2-cyclohexaloxy-4-nitrophenyl) methanesulfonamide]) (Gilroy et al., 1998) was obtained from Cayman Chemical (Ann Arbor, MI) and DFU (5,5-dimethyl-3-(3-fluorophenyl)-4-(4-methylsulfonyl)phenyl-2(5H-furanone) (Riendeau et al., 1997) was a generous gift from Dr. A. W. Ford-Hutchinson (Merck-Frosst Canada, Montreal, Canada). COX-1 inhibitor SC-560 [5-(4-chlorophenyl)-1-(4-methoxyphenyl)-3-trifluoromethylpyrazol] (Smith et al., 1998) was kindly provided by Dr. R. A. Marks (Searle, Skokie, IL). Stock solutions (10 mg/ml) were prepared in dimethylsulfoxide (DMSO):ethanol (1:4). Dilutions were made in DMSO:ethanol (1:9). Dexamethasone (dissolved in 0.9% NaCl) and indomethacin (dissolved in 2% sodium carbonate solution, and diluted with Sorensen buffer, pH 7.2) were obtained from Sigma.

Statistical Analysis. Values were calculated as means ± S.E.M. Statistical analysis was performed using Student's t test, one-way analysis of variance, or Kruskal-Wallis one-way analysis of variance, when appropriate, using Dunnett's or Dunn's post-test, respectively (SigmaStat statistical software, Jandel Scientific, Erkrath, Germany). P < 0.05 was considered statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

CT and Fluid Transfer RatesEffects of Cyclooxygenase Inhibitors and Dexamethasone. Fluid was absorbed in jejunal loops of all control rats (Figs. 1 and 2). Four hours of exposure to CT (0.5 µg/ml) caused profuse net fluid secretion (Figs. 1 and 2).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Effects of NS-398 () on CT-induced net fluid secretion and on net fluid absorption in controls (). Each point represents the mean ± S.E.M.; n in parenthesis, *P < 0.05 compared with CT .

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Effects of DFU () on CT-induced net fluid secretion and on net fluid absorption in controls (). Each point represents the mean ± S.E.M.; n in parenthesis. *P < 0.05 compared with CT.

CT-Induced PGE₂ BiosynthesisEffects of Cyclooxygenase Inhibitors and Dexamethasone. CT (0.5 µg/ml, 4 h) significantly enhanced intraluminal PGE₂ release about 3-fold compared with controls (Fig. 3). The selective COX-2 inhibitors NS-398 and DFU, which did not have any effect on PGE₂ synthesis in controls, did significantly inhibit the CT-induced increase in PGE₂ release (Fig. 3). The selective COX-1 inhibitor SC-560, as well as dexamethasone, did not influence the PGE₂ synthesis in control rats or rats treated with CT (Fig. 3).

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Effect of cholera toxin (0.5 µg/ml intraluminal 4 h) on luminal PGE2 release (pg PGE2/g/h) in the rat jejunum in the absence and presence of NS-398 (3 mg/kg), DFU (2 mg/kg), SC 560 (30 mg/kg) and dexamethasone (dexa). Each column represents the mean ± S.E.M.; n in parenthesis. *P < 0.05 compared with the effect of CT.

Effect of CT and LPS on COX-1 and COX-2 mRNA Expression in Jejunal Mucosa. COX-1 and COX-2 mRNA expression was determined using RT-PCR. COX mRNA levels were normalized to GAPDH to correct for variations in RNA concentration.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Expression of COX-2 mRNA in the rat jejunal mucosa as determined by RT-PCR. COX-2 mRNA levels were normalized to correct for variations in RNA concentration. A, ratio of the relative signal intensities COX-2/GAPDH in the jejunal mucosa exposed to tyrode (control, co) or CT (0.5 µg/ml, 4 h, ct) in rats treated with 0.9% NaCl or dexamethasone (dexa). Values are mean ± S.E.M.; n = 4 each. The insert shows representative agarose gels of the RT-PCR, lane 1 = co, lane 2 = ct, lane 3 = ct & dexa, lane 4 = dexa. The marker shows the length of the DNA in bp and the respective concentration in nanograms. B, ratios of relative signal intensities COX-2/GAPDH in the jejunal mucosa of control rats (vehicle injection, co) or rats that were treated with LPS and pretreated either with vehicle or dexamethasone (dexa). Values are means ± S.E.M., n = 3 each. The insert shows representative agarose gels of the RT-PCR, lane 1 = co, lane 2 = LPS and dexa, lane 3 = LPS .

Effect of CT on COX-2 Protein Levels. COX-2 protein levels were determined using Western blot analysis. In the supernatant fraction of the jejunal mucosa of untreated rats, the 72- to 74-kDa protein band, which specifically reacts with COX-2 polyclonal antibody, was either absent or only weakly detected (Fig. 5, lanes 1-4). All of the animals challenged with CT (0.5 µg/ml, 4 h) clearly showed a COX-2 protein band (Fig. 5, lanes 5-8), which was increased by 3- to 5-fold compared with control animals. Pretreatment of the rats with dexamethasone (1 mg/kg, twice within 24 h) did not significantly alter COX-2 protein levels in CT-treated rats (Fig. 5, lanes 9-12). In addition to these experiments performed in vivo, we have also shown that CT up-regulated COX-2 antigen levels (3- to 5-fold) in a murine macrophage cell line RAW264.7 in vitro. Untreated macrophages did not exhibit COX-2 corresponding band on Western blots (data not shown). Both for in vitro and in vivo experiments, we have used appropriate negative and positive controls in our Western blot analysis. The positive controls included LPS-treated rat mucosal intestinal tissue as well as LPS-induced macrophages, which exhibited a strong 72- to 74-kDa band. Our negative control included normal rat intestinal tissues and cell lysates from untreated macrophages, which did not react with the antibodies to COX-2.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Expression of COX-2 protein as determined by Western blot probed with COX-2-specific polyclonal antibody and developed using AP substrate kits. Equal amounts of protein (50 µg) from jejunal mucosa homogenate exposed to tyrode (controls, lane 1-4) or cholera toxin (0.5 µg/ml, 4 h, lanes 5-8) and from CT-exposed jejunal mucosa homogenate from rats pretreated with dexamethasone (lanes 9-12) were loaded onto SDS-12% polyacrylamide gel electrophoresis. Results of four different rats are given for each treatment. Three independent experiments were performed and a representative blot is shown. The blots were scanned to calculate fold increase in the antigen level using the Gel Doc 2000 system (Bio-Rad).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

It has been shown repeatedly that the release of PGs is causally involved in the secretory response to CT (for references, see Beubler et al., 1989 and Peterson et al., 1999). It is now well established that cyclooxygenase, the hemoprotein that catalyzes the formation of PGs, prostacyclins, and thromboxanes, exists in two isoforms. The concept was that the constitutive isoform COX-1 is responsible for the production of PGs involved in physiological functions, and the inducible COX-2 is expressed in response to inflammatory or mitogenic stimuli. In the meanwhile, however, several groups have reported the constitutive expression of COX-2 in several organs and tissues (see Introduction).

The present results show that CT-induced fluid secretion as well as PGE₂ formation are dose-dependently, and almost completely, inhibited by the selective COX-2 inhibitors NS-398 and DFU. The COX-1 inhibitor SC-560 did not affect either CT-induced fluid secretion or CT-induced PGE₂ synthesis. NS-398, DFU and SC-560 were used at doses previously shown to be selective for inhibition of COX-2 and COX-1, respectively (Futaki et al., 1993; Riendau et al. 1997; Smith et al., 1998; Wallace et al., 2000). Therefore it seems that CT-induced PGE₂ formation and the resulting fluid secretion are mediated primarily via COX-2. The nonselective COX inhibitor indomethacin, as shown before (Beubler et al., 1990), only partially reduces the secretory response, and it has been demonstrated previously that indomethacin inhibits PGE₂ formation (Beubler and Horina, 1990).

In this study we have shown the constitutive expression of both COX-1 and COX-2 mRNA in the rat jejunal mucosa. CT had no effect on COX-1 mRNA expression and also did not influence COX-2 mRNA expression. Treatment of rats with dexamethasone, known to interfere with COX-2 induction by transcriptional and post-transcriptional mechanisms (Newton et al., 1998), did not change the expression of COX-2 mRNA in control rats, just as it had no effect on the COX-2 mRNA expression of CT-treated rats. In contrast, LPS caused a marked induction of COX-2 mRNA expression, and as expected, this increase was totally blocked by pretreatment of rats with dexamethasone. These results suggest that COX-2 mRNA is constitutively expressed in the rat jejunum and CT did not induce COX-2 mRNA, in contrast to LPS, that is known to induce COX-2 expression in a variety of models (Herschman, 1996). A different picture was reflected when COX-2 antigen levels were analyzed by Western blotting. Although in controls, most of the samples analyzed did not show COX-2 antigen with some samples exhibiting a weak signal plausibly indicating low grade inflammation in animals, CT treatment of the jejunal mucosa caused an increase of COX-2 protein, which was not influenced by treatment of rats with dexamethasone, arguing against induction of COX-2 transcription. Studies from our laboratory and from others have indicated the specificity of the COX-2 antibodies used and have shown that these antibodies did not react with COX-1 (Muller-Decker et al., 1998, 1999; Murakami et al., 1998). Therefore, our data suggest a post-transcriptional effect of CT on COX-2. Post-transcriptional regulation of COX-2 protein has also been postulated in colon carcinogenesis (Dixon et al., 2000). It can be speculated that CT interferes with translational regulation of COX-2 or inhibits the degradation of COX-2 protein. Constitutive COX-2 expression has been shown in a variety of tissues and in this aspect, interestingly, MacNaughton et al. (2000) have demonstrated that COX-2 is constitutively expressed in the mouse colon. The finding that dexamethasone did not affect the CT-induced PGE₂ formation and fluid secretion further substantiates the concept of a constitutive expression of COX-2. However, since we used total mucosal scrapings of the jejunum, the results concerning mRNA and protein are averaged and we cannot exclude that changes in mRNA and/or protein levels may occur in discrete subpopulations of cells. The mechanism of CT-induced PGE₂ synthesis is poorly understood. However, it is known that CT increases phospholipase A₂ activity, probably by induction of phospholipase A₂-activating protein mRNA (Peterson et al., 1996). Phospholipase A₂ specifically hydrolizes fatty acids from phospholipids and constitutes a central point of control for PG biosynthesis. The present data suggest that arachidonic acid is metabolized via COX-2 to PGE₂ and that selective inhibition of COX-2 significantly inhibits CT-induced PGE₂ biosynthesis and that results in reduced fluid secretion.

The isoform of cyclooxygenase, COX-1 seems not to be involved in the mediation of CT-induced fluid secretion. The specific COX-1 inhibitor SC-560 did not affect CT- induced fluid secretion and PGE₂ formation, despite near maximal inhibition of platelet thromboxane B₂ biosynthesis (data not shown). Indomethacin, which inhibits both COX-1 and COX-2, only partially reduces the secretory response, as shown before (Beubler and Horina, 1990), although via its COX-2 effect, complete inhibition would have been expected. Also in human cholera, results obtained with indomethacin are not so clear. In experiments performed in Bangladesh (Rabbani and Butler, 1985), indomethacin was not able to inhibit fluid loss; however, Van Loon et al. (1992) described a significant inhibition of PGE₂ and net fluid secretion in humans with indomethacin.

It may be speculated that indomethacin, due to its inhibitory effect on COX-1, is more toxic at the dose used than the pure COX-2-inhibiting compounds NS-398 and DFU. Although indomethacin does inhibit PGE₂ overflow into the gut lumen like NS-398 and DFU, this does not reflect inhibition of PGE₂ formation in the tissue. In the case of indomethacin, this inhibition may be incomplete, as already stated by Rabbani and Butler (1985).

Similar to our data, NS-398 significantly blocked PGE₂ formation and the secretory response induced by arachidonic acid in mouse colon, suggesting that a substantial portion of arachidonic acid-induced secretion was mediated through a COX-2-dependent mechanism, whereas the response to arachidonic acid was also blocked by pretreatment with the selective COX-1 inhibitor SC-560, suggesting the involvement of both isoenzymes in this kind of stimulation (MacNaughton et al., 2000). In contrast, CT seems to operate selectively through COX-2.

Since arachidonic acid uses both isoenzymes, and CT operates only via COX-2, it may be speculated that COX isoenzymes are activated in specific intracellular locations or in specific cell types to produce the final common secretagogue, PGE₂.

There is no doubt that the enteric nervous system is also involved in the mediation of CT- or prostaglandin-induced secretion (Cassuto et al., 1981; Jodal et al., 1993). CT may bind directly to neurons in the myenteric plexus (Jiang et al., 1993) or stimulate neurons via 5-HT₃ receptors after having released 5-HT from enterochromaffin cells (Beubler and Horina, 1990). An important role is also played by 5-HT in the mediation of CT-induced mucin secretion, using primarily 5-HT₄ receptors (Moore et al., 1996). On the other hand, 5-HT binds to receptors on cells of the mucosa, presumably 5-HT₂ receptors (Beubler and Horina 1990), leading to the initiation of arachidonic acid metabolism and release of PGs (Beubler et al., 1989; Siriwardena et al., 1993). To summarize these observations, it can be stated that CT uses 5-HT, PGs, the enteric nervous system, and possibly also cyclic AMP to exert its effect; however, the exact nature of the cascade of events leading to CT-induced fluid secretion remains speculative.

In conclusion, it has been demonstrated that CT, besides other mediators, uses PGE₂ to exert its tremendous secretory effect. In the case of CT, the PGs released are solely produced via COX-2. Since no induction of COX-2 mRNA was observed in total mucosal scrapings of the jejunum, and dexamethasone was without an effect on COX-2 mRNA or COX-2 protein levels, it can be assumed that post-transcriptional modulation is responsible for the CT-induced increase in COX-2 protein levels. COX-1 seems not to be involved in CT-induced fluid secretion, a finding that distinguishes COX-2 as a new and specific target for the treatment of Asiatic cholera.